Exchange 2007 / 2010 environment. I have setup a CAS array using two separate servers. BOth of these servers have CAS and HUB ONLY installed. They do not have the mailbox role installed. I went through the steps of setting up the CAS Array and both of these servers see each other just fine. I assigned a pingable, reachable ip address to the array. I added the name cashub.mydomain.com to our internal DNS database. I went to one of these servers and ran these commands: New-ClientAccessArray -Name cashub -Site "Default-First-Site-Name" -Fqdn cashub.mydomain.com Get-MailboxDatabase -Server maila | Set-MailboxDatabase -RpcClientAccessServer cashub.mydomain.com I verified that the new RPC value shows up as cashub.mydomain.com on the server. I then closed outlook and went to MAIL under control panel to modify my outlook profile. My mailbox was already moved to the Exchange 2010 server, so I did a profile repair. That should update my profile to see that it should use cashub.mydomain.com from now on and not the original single server name. I verified that the change did in fact take place. However, when I open outlook, it pauses abnormally (15 seconds), pops up a Login screen pointing to cashub.mydomain.com with mydomain.com\myname in the username box. I type in my password and then outlook freezes for a while. I eventually get an error reading: Cannot start Microsoft Office Outlook. CAnnot open the Outlook window. The set of folder cannot be opened. Microsoft Exchange is not available. Either there are networ problems or the Exchange computer is down for maintenance. If I change the RPCClient AccessServer back to one of the single CAS/HUBS, Outlook works just fine, like it always did. Is there something that I've missed? Why would the cas array ask for credentials when the individual cas/hub servers authenticate without asking for credentials? Any ideas or help is greatly appreciated. Thanks in advance.

What does the IP address point to? One of the servers, an NLB? I always start by pointing the CAS array DNS name at one of the CAS server's IP addresses. That allows me to confirm it works before I move forward to anything more sophisticated. You also need to ensure the name resolution works correctly as well and that the name doesn't resolve externally. Simon.Simon Butler, Exchange MVP Blog | Exchange Resources

Thanks for your response. The ip address points to the array IP address (NLB) which is different from the two cas addresses. So, cas1 = 10.x.x.1, cas2 = 10.x.x.2, cashub array = 10.x.x.3. DNS entry for cashub.mydomain.com = 10.x.x.3. Also, I am using dual NICs. I am told that a single NIC will work as well, but that using Dual NICs is better as it splits up the traffic. Name resolution does work, but, I will remove the array, re-construct it and use only one cas server to test it out. Also

I just reconstructed the NLB, using only one of the CAS servers (the one that we were using before I setup an array). I see the same behavior as before. After repairing my Outlook 2007 profile to point to the cashub array, I am prompted for my password and Outlook never connects. When I go to the connection status on the Outlook system tray icon, I can see it trying to get to the CASHUB array, but the status reads connecting. It never shows established. From this same PC, if I do an nslookup, it resolves to the correct array address, and I can ping the cas array address. I'm lost as to what can be causing this to happen.

You don't need the NLB for a single server test. Point the CAS array DNS entry directly at the CAS server and test again. You need to prove the CAS array alone is working correctly, without introducing the additional complexity and potential issues of the NLB . Simon.Simon Butler, Exchange MVP Blog | Exchange Resources

I guess I misunderstood. If you are saying to change the DNS entry for the CASHUB array to point to the IP of a single CAS server, I just did that and it works fine. The connection status shows my Outlook talking to CASHUB, all with the status of established. I don't believe this to be a DNS issue. It seems more like an issue of traffic not flowing through the array when NLB is setup. All servers are on the same subnet so I'm not sure why there would be any routing issues.

That firmly points the finger at NLB. Nothing to do with Exchange, something is wrong with NLB. That is kind of what I expecting. There are known issues with NLB and certain NICs and network switches. The switch cannot handle the "virtual" NIC correctly. Simon.Simon Butler, Exchange MVP Blog | Exchange Resources

So, is it worth trying the single NIC implementation of NLB? I've heard that may be an option. It seems multicast would need to be used.

That could be one way to work. You might also want to look around the Internet to see if there are known issues with your combination of switch, NIC and NLB. If you are using VMs, then don't forget to throw that in to the mix as well. Simon.Simon Butler, Exchange MVP Blog | Exchange Resources